ARTICLE IN PRESS

Tunnelling and Underground Space Technology xxx Ž2001. xxx᎐xxx

A study of liquefaction related damages on shield tunnels H.S. Choua,U , C.Y. Yang a , B.J. Hsieh a , S.S. Chang b b

a CTCI Corporation, Taipei, Taiwan, ROC Department of Rapid Transit System, Taipei City Go¨ ernment, Taipei, Taiwan, ROC

Received 15 April 2001; received in revised form 20 August 2001; accepted 23 August 2001

Abstract Studies of liquefaction-related damages on underground structures are limited and sometimes controversial. Liquefaction potential analysis is essential in tunnel design in liquefiable soils. A Taipei Rapid Transit System ŽTRTS. tunnel site in Taipei County was selected to study the risk under liquefaction-related damage. The liquefaction risk index was applied for assessment of the overall liquefaction risks and liquefaction-induced settlement at the interest site. The anti-liquefaction measure for shield tunneling by using secondary injection grouting was discussed to eliminate the flow of liquefied soils and reduce the liquefaction-induced settlement. 䊚 2001 Published by Elsevier Science Ltd. Keywords: Liquefaction and liquefied soils; Uplift; Shield tunnel; Liquefaction-induced settlement; Secondary injection grouting

1. Introduction Taiwan is located within the mid-section of circumpacific seismic zone, where the Philippine Plate plunges into the Eurasian Plate, and has been experiencing extensive seismic activities. During severe earthquakes, a loose saturated sand deposit may lose its strength temporarily if the progressively build-up excess pore water pressure induced by the earthquake approaching the initial effective confining pressure of soil. This phenomenon is so-called soil liquefaction. The possible damages induced by liquefaction include floating or sinking of underground structure, reduction of soil bearing capacity, and increase of lateral pressure. After the Chi-Chi earthquake in Taiwan on 21 September 1999, many above ground structures such as bridge, road and buildings suffered from the liquefaction-related damages and led to tilt and overturning U

Corresponding author. CTCI Corporation, 77 Sector 2 Tun Hwa South Road, Taipei, Taiwan, ROC. Tel.: q886-2-2700-9659; fax: q886-2-2709-4091. E-mail address: [email protected] ŽH.S. Chou..

failure. Some underground structures have also been damaged ŽWang et al., 2001.. In general, liquefactionrelated damages can be divided into three categories including superstructure damage, liquefaction-induced settlement, and underground structure damage. The first two categories of damages are easy to detect since they occur on ground surface. Yet, studies of liquefaction-related damage on underground structure such as shield tunnels and common ducts are limited and often controversial due to lack of physical evidence and difficulty in investigation. In addition, liquefaction-related damage on underground structures is less significant than that on above ground structures. It may be the reason that underground structures such as shield tunnels have flexibility characteristics compatible with the surrounding soils. During strong motion, the flexible structures can deform compatibly with the surrounding soils and are then less likely to induce a rigid type of failure. Shield tunnels are generally at depth with certain thickness of overburden Žapprox. 10 m or more.. It is noted that the liquefaction-induced settlement is contributed mainly by the soil layer near the surface and thus may have little influence on the shield tunnel.

0886-7798r01r$ - see front matter 䊚 2001 Published by Elsevier Science Ltd. PII: S 0 8 8 6 - 7 7 9 8 Ž 0 1 . 0 0 0 5 7 - 8

ARTICLE IN PRESS 2

H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

Moreover, ground treatment measures are usually applied for building protection when the shield tunnel passes through the soft or loose soils. Consequently, the shear strength of the treated soils is increased, which leads to a much lower liquefaction potential. Considering the importance of underground public facilities such as the mass rapid transit system, safety is the major concern. Liquefaction potential analysis is essential when the shield tunnel is located at the liquefied layer. A Taipei Rapid Transit System ŽTRTS. tunnel site in Taipei County of Taiwan was selected to study its risk under liquefaction-related damage. The liquefaction risk index proposed by Iwasaki et al. Ž1982. was applied for the overall assessment on liquefaction risks at the site. The concept suggested by the Design Guideline for Common Duct ŽJapan Road Associate, 1986. was introduced for uplift analysis under the soilliquefied state. In addition, the methods proposed by Tokimatsu and Seed Ž1987. and Ishihara and Yoshimine Ž1992. were adopted for the evaluation of liquefactioninduced settlement. Based on those analyses including the liquefaction potential, soil properties, impacts on traffic and environment, and cost and schedule, antiliquefaction measures by using secondary injection grouting are applied to eliminate the flow of liquefied soils and to reduce liquefaction-induced settlement.

2. Engineering geology of the site The study site is located in the west side of the Taipei Basin. The stratigrphy shows that the Quaternary sedimentary deposits are overlying the Tertiary bedrock. the Quaternary deposits can be further classified into three major formations: the Hsingchuang Formation on the bottom; the Chingmei Formation in the middle; and the Sungshan Formation on the top. The Sungshan Formation is the most recent deposit of 40᎐70 m thick, consisting of alternative layers of silty clay and silts interstratified with fine sand layers with high silt content. The Sungshan Formation is of prime engineering concern because of its high compressibility and low shear strength. Hung Ž1966. proposed that the subsoils of the Sungshan Formation can be subdivided into six sub-layers. The fifth, third and first sub-layers of the Sunshan Formation are mainly sandy soils, whilst sixth, fourth and second sub-layers are silty and clayey soils. According to eight field boring results from B34᎐B-41, the soil condition of the site is shown in Fig. 1. Since the fifth sub-layer is mainly composed of silty sand with loose to medium dense, the sand formation will be susceptible to earthquake-induced liquefaction. The depth of the shield tunnel center is between 14 and 16.5 m below the ground surface. The tunnel has

Fig. 1. Soil profile along the interest site.

ARTICLE IN PRESS H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

3

Fig. 2. Related soil parameters varied with depth.

an inner diameter of 5.6 m and an outer diameter of 6.1 m. The tunnel is located in the Sungshan Formation. For the analysis of liquefaction potential, the related soil parameters which varied in depth were summarized and presented in Fig. 2.

Road Associate Ž1996. and Design Specification of Taiwan’s Building Code Ž1997.. Some relevant parameters of the study site were adopted as below: 䢇 䢇

3. Analysis of liquefaction potential

䢇

In Taiwan, the commonly used methodology for liquefaction evaluation in engineering practice was the semi-empirical stress-comparison method based on SPT N-value. Four approaches were applied to analyze the liquefaction potential of the study site, including Tokimatsu and Yoshimi Ž1983., Seed et al. Ž1985., Japan

䢇 䢇

Earthquake magnitude: 7.1. Peak ground acceleration at horizontal direction for design: 0.23 g. Average elevation of ground surface: EL.q 110.63 m. Ground water level: EL.q 109.63 m. Energy ratio of standard penetration test: 60%.

Based on the eight boring data shown in Fig. 1, the liquefaction potential at the site was evaluated by using the above four different approaches. The factor of

Table 1 Results of liquefaction risk index Boring No.

The liquefaction risk index PL Taiwan’s Building Code Ž1997.

Seed et al. Ž1985.

Japan Road Associate Ž1996.

Tokimatsu and Yoshimi Ž1983.

B-34 B-35 B-36 B-37 B-38 B-39 B-40 B-41

2.94 1.72 14.86 14.26 11.72 8.10 3.25 2.84

0.00 2.07 18.31 14.02 2.89 7.52 0.94 2.94

6.62 2.30 19.22 23.52 16.85 12.55 3.13 5.78

4.20 1.57 13.60 16.44 14.78 6.33 1.51 3.96

Ave.

7.46

6.08

11.25

7.80

0 F PL - 5 s unlikely to slightly; 5 F PL - 15 s slightly to medium; PL ) 15 s likely.

ARTICLE IN PRESS 4

H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

Fig. 3. Results of analysis of liquefaction potential.

safety against liquefaction Ž FL . was plotted in Fig. 3. In addition, the liquefaction risk index Ž PL . proposed by Iwasaki et al. Ž1982. was applied to evaluate the comprehensive risk of the site under soil liquefaction. This index is determined based on the thickness and depth of liquefied soil and its corresponding safety factor. The liquefaction risk indexes Ž PL . estimated from eight boreholes are summarized in Table 1. The findings incorporated with the factor of safety against liquefaction Ž FL . are outlined below:

4. Liquefaction-induced uplift

1. The study of liquefaction potential is focused on the subsoil at the depth of the tunnel and its bottom layer. 2. Results from the analyses by four different methods indicate that the subsoil near the tunnel is susceptible to earthquake-induced liquefaction at the section B-39᎐B-40 ŽFig. 3.. The liquefiable soil is of average thickness of 2᎐3 m. 3. All the analyses, except the analysis by the Seed et al. Ž1985. method, show that subsoil Žsection B35᎐B-37. close to the tunnel bottom is prone to liquefaction under the designed earthquake ŽFig. 3.. The liquefiable soil layer below the shield tunnel is of average thickness of 3 m. 4. For further understanding the risk on liquefaction-related damage, liquefaction risk index Ž PL . is applied. As shown in Table 1, the tunnel at section B-35᎐B-37 and B-39᎐B-40 will expect a light to medium degree of damage under the attack of soil liquefaction. 5. In order to determine the most susceptible area at the study site, an overall assessment is performed in accordance with all the analyses including the liquefaction risk index of each borehole, soil profile

During earthquake shaking, the built-up excess pore water pressure will lead to the reduction of the effective stress of soil and the soil will progressively approach the liquefaction state. The possible damages induced by liquefaction include floating or sinking of the underground structure, reduction of soil bearing capacity and increase of lateral pressure. Based on the investigation in Japan’s underground structure, the line-type structures such as tunnels and common ducts contributed few cases of earthquake-related damages. It is recommended that liquefaction-induced uplift should be examined for the design of tunnels and common ducts. Liquefaction-induced settlement should also be considered. In general, the design guideline of the common duct suggested by Japan Road Associate Ž1986. can be adopted for checking the liquefaction-induced uplift. The methodology is based on the resistance contributed from friction between the underground structure and surrounding soils, the factor of safety against liquefaction Ž FL ., and the driven force considering the built-up excess pore water pressure during soil liquefaction. According to the results of liquefaction analysis, the

and distribution of SPT N value and the relevant safety factor at a different location and depth, as presented in Fig. 4. The study indicates that the sandy soil underlying the shield tunnel has the most susceptible impact on liquefaction-related damage at the site.

ARTICLE IN PRESS H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

5

Fig. 4. Overall assessment of liquefaction potential Žfactor of safety resulting from average value of four methods..

factor of safety against liquefaction Ž FL . at the depth close and underneath the tunnel bottom is mainly between 0.8 and 0.9. Besides, the liquefaction risk index Ž PL . analysis indicates that the site can be classified as a light to medium degree of risk for the liquefaction related damage. Borehole B-39 was selected for the checking of liquefaction-induced uplift. At this location, the ground is with much less overburden and the calculated factor of safety against liquefaction is 0.81, and the calculated safety factor against uplift is 1.08, which is greater than the value specified in design code Ž1.07.. In other words, the possibility of liquefaction-induced uplift failure of the tunnel at this location is low.

5. Liquefaction-induced settlement The post-earthquake densification of saturated sand is influenced by the density of sand, the maximum shear strain induced in the sand and the amount of excess pore pressure generated by the earthquake. Laboratory experiments have shown that the volumet-

ric strain after initial liquefaction varies with relative density and maximum shear strain. There are two approaches proposed by Tokimatsu and Seed Ž1987. and Ishihara and Yoshimine Ž1992. to estimate the settlement of subsoil induced by liquefaction. Tokimatsu and Seed Ž1987. used a correlation between Ž N1 . 60 and cyclic stress ratio ŽCSR. to produce a chart, as shown in Fig. 5. It allows the volumetric strain after liquefaction in an earthquake of Ms 7.5 to be estimated directly from the CSR and SPT resistance. In an alternative approach by Ishihara and Yoshimine Ž1992., the factor of safety against liquefaction, the maximum cyclic shear strain, the relative density, the SPT resistance N1 w( 0.833Ž N1 . 60 x, or the CPT tip resistance, can be used to estimate post-earthquake volumetric strain, as shown in Fig. 6. One case study based on the soil data of B-39 at the study site was selected. The method proposed by Seed et al. Ž1985. was adopted to evaluate the cyclic stress ratio ŽCSR. and factor of safety against liquefaction of soils, based on the volumetric strains over the thickness of the liquefied layer that produces the liquefaction-induced settlement. The analyzed results are summarized

ARTICLE IN PRESS H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

6

Fig. 5. Relationship between volumetric strain Ž ␧ v ., cyclic stress ratio ŽCSR. and Ž N1 . 60 Žafter Tokimatsu and Seed, 1987..

in Table 2. It indicates that the settlements estimated by Tokimatsu and Seed Ž1987. at ground surface and tunnel bottom are 23 and 9 cm, respectively, while by Ishihara and Yoshimine Ž1992. are 34 and 13 cm, respectively. The settlement of the liquefied layer underneath the tunnel generated by the post-earthquake densification of saturated sand is substantial and should be carefully considered.

Fig. 6. Relationship between volumetric strain Ž ␧ v . and factor of safety against liquefaction Ž FL . Žafter Ishihara and Yoshimine, 1992..

induced settlement is expected, soils surrounding the tunnel require improvement. Soil improvement techniques may be adopted to: 䢇

䢇

6. Anti-liquefaction measures for shield tunnel

䢇

6.1. Assessment of ground treatment method The analyses in the previous sections indicated that local liquefaction is possible at the site of loose to medium dense silty-fined sand. Since the liquefaction-

provide protection against liquefaction on supporting soils underneath the tunnel; reduce liquefaction-induced settlement; and minimize the post-earthquake flow of liquefied layer.

Liquefaction is triggered because the soil is loose and the excess pore water pressure cannot dissipate rapidly. Thus, measures against liquefaction are either to increase soil density or to improve its drainage path. Soil improvement methods including stone column,

Table 2 Anticipation of liquefaction-induced settlement Depth Žm.

N

9.0 10.5 12.0 13.5 15.0 18.0 19.5

11 13 12 37 13 10 11

Ž N1 .60

11.6 12.7 11.1 32.6 11 7.8 8.2

N1

9.7 10.6 9.2 27.2 9.2 6.5 6.8

FL

0.76 0.79 0.72 2.31 0.87 0.73 0.84

Tokimatsu and Seed Ž1987.

Ishihara and Yoshimine Ž1992.

␧v Ž%.

␧v Ž%.

2.3 2.2 2.4 ᎐ 2.4 3 2.9 Total settlement at ground surface Ý H s Total settlement at the bottom of tunnel Ý H s

⌬ H Žcm. 3.45 3.3 3.6 ᎐ 3.6 4.5 4.35

22.8

8.85

3.5 3.3 3.6 ᎐ 3.6 4.3 4.2 Total settlement at ground surface Ý H s Total settlement at the bottom of tunnel Ý H s

⌬ H Žcm. 5.25 4.95 5.4 ᎐ 5.4 6.45 6.3

33.75

12.75

Based on the soil data of B-39, the method of Seed et al. Ž1985. was adopted. The bottom of the tunnel is approximately 17 m.

ARTICLE IN PRESS H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

7

Table 3 Assessment of soil improvement method Improvement method

Stone column

Compaction grouting

High-pressure injection grouting ŽJSP.

Secondary injection grouting

Improvement area

Route at the site ␾80 [email protected] m

Route at the site ␾80 [email protected] m

Route at the site ␾80 [email protected] m

Below the springline of tunnel, with 3-m length extended from segment

Cost ŽNT. Traffic Environment Schedule Vibration

41 100 000 Serious Serious Serious Serious

85 700 000 Serious Serious Serious Medium

25 300 000 Slightly Medium Medium Slightly

10 100 000 None None None None

compaction grouting, high-pressure injection grouting ŽJSP. and secondary injection grouting Ždouble-tube grouting. were considered for the study site. In order to select the best option, assessment is made based on construction cost and schedule, impacts on environment and traffic, and utility relocation, as shown in Table 3. The comprehensive assessment indicated that soil improvement by the secondary injection grouting technique is the most appropriate measure for the site. Since secondary injection grouting is an underground activity and is executed inside the tunnel, less or no impact on traffic, environment and utility relocation are expected. In addition, secondary injection grouting can be performed through the spared grouting hole installed in the segmental ring while the shield tunnel can be advanced at the same time. The influence on construction schedule by the secondary injection grouting is minimum. The construction cost related to the secondary injection grouting is also lower compared with that by other options. Consequently, the secondary injection grouting of the surrounding soils below the springline of the tunnel, with 3-m length extended from the segment was applied at every two rings. It increased the shear strength of the surrounding soils and reduced the liquefaction-induced settlement, and at the same time, kept a certain degree of flexibility for the shield tunnel. The design of the secondary injection grouting is illustrated in Fig. 7. Grouting was applied along the tunnel route for a length of approximately 180 m at the site, as shown in Fig. 4.

and permeation procedure is introduced. The flowable particulate grout by cement and bentonite is firstly intruded into voids and cracks to cement and strengthen the surrounding subsoil. Properly formulated silicabased chemicals are then injected and permeated into sandy soils to increase the shear strength and durability of the sand, and to stabilize the loose sands against liquefaction. The characteristics of treated soil after the secondary injection can be improved substantially, e.g. the cohesion of a sitly soil in a subway of Tokyo increased from 2.76 to 6.60 trm2 , and the reaction modulus increased by 2.8 times. In the restoration of the liquefaction after the Kobe earthquake in 1995, the dual-pipe and double-packer multiple grouting was adopted with an injection ratio of 30% and strength of treated soils reached 5 kgrcm2 . The applications of secondary injection grouting in shield tunneling including three projects in Japan and six projects in Taipei are studied. The information summarizing the soil condition, injection planning and grout properties are presented in Table 4. The technique of the secondary injection grout has been widely

6.2. Secondary injection grouting Secondary injection grouting applied for the shield tunneling is the technique by using the dual-pipe and double-packer multiple grouting, in which the intrusion

Fig. 7. Sketch of secondary injection grouting.

ARTICLE IN PRESS 8

H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

Table 4 Case collections of secondary injection grouting Case

Management

Soil condition

Injection material

No. 8 Subway in Tokyo, Japan Water supply pipe in Tokyo, Japan Tokyo-Chiba route of subway, Japan

Injection pressure: static earth pressure q 0.5 kgrcm2 Injection pressure: 8 kgrcm2 Injection pressure: 5 kgrcm2

Soft clay

Suspension type: cement based

Sandy silt

Contract CH221 of Taipei RTS

Injection pressure: 5 kgrcm2 Rate of discharge: 300 lrcycle Injection pressure: initial injection pressure applied up to 2 kgrcm2 Injection pressure: initial injection pressure applied up to 2 kgrcm2 Injection pressure: 5 kgrcm2 Injection pressure: 5 kgrcm2 Injection pressure: 5 kgrcm2

Sandy silt

Solution type: inorganic reagent Semi-suspension type: cement based q compound Solution type: inorganic reagent

Contract CH218 of Taipei RTS

Contract CP261 of Taipei RTS

Contract CP264 of Taipei RTS Contract CC276 of Taipei RTS Contract CC277 of Taipei RTS

used and well developed in Taiwan. The materials for injection can be mainly divided into three types, including solution type, semi-suspension type and suspension type. The cement-based suspension grout is the most popular type, and is confirmed mostly by the injection pressure. Depending on the overburden stress of treated soils, the injection pressure is approximately 5 kgrcm2 . The injection ratio applied for Contract CP261 of the subway of Taipei is 3%. The soil condition is mainly composed of silty sand, the same as the conditions of the study site.

7. Conclusions and recommendations Based on the analyses of liquefaction potential under horizontal peak ground acceleration of 0.23 g, a liquefiable layer with a thickness of 3 m underlying the shield tunnel at the section between B-34 and B-41 is identified. An overall assessment by using the liquefaction risk index was introduced to determine its influence. One location in the B-39 borehole was selected to examine the potential uplift induced by excess pore water pressure and the liquefaction-induced settlement. The finding in B-39 shows that the factor of safety against uplift can reach the design limit. Anticipated liquefaction-induced settlement is 23᎐34 cm at

Soft clay

Silty sand or silty clay

Solution type: inorganic reagent

Silty sand

Solution type: inorganic reagent

Silty sand

Suspension type: cement based Suspension type: cement based Suspension type: cement based

Silty sand or silty clay Silty clay

ground surface and 9᎐13 cm at the bottom of the tunnel, respectively. The anti-liquefaction measures at the site are mainly focusing on the issues related to the reduction of the flow of liquefied soil and liquefactioninduced settlement. After the overall assessment, soil improvement by using secondary injection grout Ždual-pipe and doublepacker multiple grout. of the surrounding soils under springline of the tunnel was carried out. The grouting was conducted at every two rings with an extended length of 3 m. Cement-based grout material of suspension type with a 3% injection ratio and a 5-kgrcm2 injection pressure was adopted. The soil improvement measure increases both the strength and stiffness of the soils, and at the same time, reduces the liquefaction-induced settlement. Studies of liquefaction-related damages on underground structures are limited and sometimes controversial. The liquefaction potential analysis is essential and important for public underground structures such as mass rapid transit tunnels in the liquefied soils. This paper presents a study of liquefaction effects on shield tunnels using the existing methods, and presents a proposal of feasible measures against liquefaction for a specific site. Further investigation is recommended to verify these methods and to develop appropriate methods for general applications.

ARTICLE IN PRESS H.S. Chou et al. r Tunnelling and Underground Space Technology xxx (2001) xxx᎐xxx

Uncited references Hwang and Chen, 1998; Hwang et al., 1998; Kramer, 1996; Lai et al., 1997 References Design Specification of Taiwan’s Building Code, 1997. Hung, J.J., 1966. Physical properties of quaternary sediments in Taipei Basin. Eng. J. Natl. Taiwan Univ. 10. Hwang, J.H., Chen, C.H., 1998. Evaluation of design guideline on soil liquefaction assessment. Sino-Geotechnics Ž70., pp. 23᎐44. Hwang, J.H., Lee, C.J., Hwang, F.K., 1998. Study on Hsinhua Fault and Liquefaction Potential Analysis at South 2 nd Highway. Final Report for National Expressway Engineering Bureau of Ministry of Traffic and Communication. Ishihara, K., Yoshimine, M., 1992. Evaluation of settlements in sand deposits following liquefaction during earthquake. Soils Found. 32 Ž1., 29᎐44. Iwasaki, T., Arakawa, T., Tokida, K., 1982. Simplified Procedures for Assessing Soil Liquefaction During Earthquakes. Proc. Confer-

9

ence on Soil Dynamics and Earthquake Engineering. Southampton, pp. 925᎐939. Japan Road Associate, 1986. Design Guideline for Common Duct. Japan Road Associate, 1996. Specifications for Highway Bridges, vol. V. Earthquake Resistant Design. Kramer, S.L., 1996. Geotechnical Earthquake Engineering. Lai, C.H., Yu, M.S., Wu, C.M., Wu, W.K., 1997. Reduce the ground settlement by using the secondary grout injection in shield tunnelling. Sino-Geotechnics Ž60., pp. 83᎐96. Seed, H.B., Tokimatsu, K., Harder, L.F., Chung, R.M., 1985. The influence of SPT procedures in soil liquefaction resistance evaluation. J. Geotech. Eng., ASCE 111 Ž12., 1425᎐1445. Tokimatsu, K., Seed, H.B., 1987. Evaluation of settlements in sand due to earthquake shaking. J. Geotech. Eng., ASCE 113 Ž8., 861᎐878. Tokimatsu, K., Yoshimi, Y., 1983. Empirical correlation of soil liquefaction based on SPT N-Value and fines contents. Soils Found., JSSMFE 23 Ž4., 56᎐74. Wang, W.L., Wang, T.T., Su, J.J., Lin, C.H., Seng, C.R., Huang, T.H., 2001. Assessment of damages in mountain tunnels due to the Taiwan Chi-Chi earthquake. Tunnelling Underground Space Technol. 16 Ž3..